Let us pause for a moment to compare this with an alternative hypothesis, which some astronomers have favoured, that stars are being created all the time. On this hypothesis we picture the stars as passing in an endless steady stream from creation to extinction, just as men pass in an endless steady stream from their cradles to their graves, a new generation always coming into being to step into the place vacated by the old. On this view Plaskett’s star, with about a hundred times the weight of the sun, must be a recent creation, while Kruger 60, with only a fraction of the sun’s weight, would be very, very old—perhaps 100 million million years older than Plaskett’s star. At present direct observation cannot definitely decide between the two conflicting hypotheses, but it rather frowns upon the “steady stream” view of the stars. In a steady population the number of people in any assigned condition is exactly proportional to the time taken to pass through that condition. Suppose for instance that human beings possess infant teeth for a quarter as long as they possess adult teeth. If examination of the teeth of a population shewed that four times as many had adult teeth as infant teeth, this would create a _prima facie_ expectation that we were dealing with a steady population. If, on the contrary, 100 times as many people were found with adult teeth as with infant teeth, we should know we were not dealing with a steady population. If other evidence pointed to the population all being of approximately the same age, we should be inclined to accept this and regard the 1 per cent. of cases of infant teeth as cases of arrested development. We do not judge the ages of stars by their teeth but by their weights and luminosities. And the luminosities of the stars are not found to conform to the statistical laws which would prevail in a steady population of stars. There appear to be so many middle-aged stars and so few infants and veterans as to make the hypothesis of a steady continuous creation hardly tenable. Indeed there is rather distinct evidence of a special creation of stars at about the time our sun was born. This leads back again quite naturally to the view that the galactic system was born out of a spiral nebula whose main activity as a parent of stars occurred some 5 to 10 million million years ago.